Hi EbNauters
I have an SL instance running but not surprise to get no decode here. Too much noise :(
Anyway would be interesting to discuss the hardware needed for decoding EbNaut in LF
Would help others to figure out how to start using their stations already decoding WSPR, etc
I'm using Perseus SDR sampling 500KHz with SDR-Radio V2.3 from Simon Brown
SDR-Radio can set up to six different VFO's in those 500KHz and send the audio of each one
to different Virtual Audio channels. So it is like having 6 different receivers covering VLF, LF and MF
There is one VFO at 136KHz demodulating USB at 4KHz wideband. This audio is sent to a VAC channel
to feed WSPR-2, Opera and two instances of Spectrum Lab, one for showing spectrogram in my grabber
and the new one for EbNaut to generate FFT files and then process with the decoder
First question: There is no sample rate correction here as I'm using virtual audio cable and not
a hardware sound card. Is that a correct asumption ?
So the only problem about frequency accuracy should be the SDR receiver clock. Right ?
SpectrumLab instance for EbNaut is getting the virtual cable audio channel and doing a complex FFT
with internal frequency shift at 1395Hz. As the dial frequency at SDR-Radio is 136000Hz
then we got the center frequency at 136000 + 1395 = 137395 Hz just at Rob's frequency
The FFT decimation is 1024 and FFT lenght 131072 which produces
Width of one FFT-bin: 328.571 uHz
Equiv. noise bandwidth: 328.571 uHz
Max freq range: 1.38423 kHz .. 1.40577 kHz
FFT window time (length): 50.72 min
The File Export of FFT is activated. So, a new file is created every 50.72 min and contains 50.72 min
of signal recorded. As transmission starts at 00 and 30 min past the hour, every file should contain
a complete message with enought margin to fit it
Second question: Is that correct so far ?
If an analog receiver is used (non SDR) then one should dial 136KHz USB and route the audio to PC sound card
I guess there would be two sources of frequency error that way. One would be the Rx own drift and precision
and the other would be the soundcard sample rate error
Sound card sample rate can be determined and compensated by SpectrumLab just by injecting 1pps signal from GPS
to the right channel of sound card (having signal in left channel). And for the Rx drift and precision that depends a lot
on the receiver technology as you already know
I guess than Rx being stable enought, then you just need to determine your Rx shift error and then can be adjusted
later at decoding stage by configuring the decoder to watch for the signal at the right place
Please confirm if all this sounds phase coherent ;-)
73 de Luis
EA5DOM
